predictive maintenance from system integration perspective · manfred austen ceo predictive...

21
Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential Future Solutions 19.01.2017

Upload: others

Post on 09-Aug-2020

6 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

Manfred Austen

CEO

Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential Future Solutions

19.01.2017

Page 2: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17

• (1) Challenge Maintenance – Definition and Strategies of Maintenance – Problem and Objectives of Maintenance

• (2) Current Situation – Prerequisites and Aspects – Classic Manufacturing IT Systems – Corrective and Periodic Maintenance – Typical Business Models

• (3) Impact of IoT, Cloud Computing and Big Data – IoT Boxes, Cloud Computing, Big Data and their potential – “Trendy “Manufacturing IT Systems – Condition Based Maintenance – Upcoming Business Models

• (4) Potential Future Solutions ? – On the Edge Computing & Machine Learning – “Future” Manufacturing IT Systems – Predictive Maintenance Approaches – Future Business Models

• (5) Conclusion

2

Predictive Maintenance from System Integration Perspective

Agenda

• At one of SYSTEMA’s customers after each shift the operators fill out a form to inform the maintenance team about problems that have occurred and which require repair or correction.

• In exchange, the maintenance team informs the operators about the measures they have taken.

• In this presentation are some complaints and problems, which were actually submitted, together with the respective response of the maintenance team.

• One can not claim that neither operators nor maintenance team members were humorless.

• Explanation P = Problem reported S = Solution applied

• Extract from the Log Book see following slides

Page 3: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17

• Maintenance, Repair, and Operations (MRO) [Wikipedia] – Involves fixing any sort of mechanical, plumbing, or electrical

device should it become out of order or broken. – It also includes performing routine actions which keep the

device in working order or prevent trouble from arising. – Refers to any maintenance, repair and operation activity to keep

a manufacturing plant running.

• Plant Maintenance – Routines and recurring work required to keep a facility in such

condition that it may be continuously used, at its original or designed capacity and efficiency for its intended purpose.

– Actions necessary for retaining or restoring a piece of equipment, machine, or system to the specified operable condition to achieve its maximum useful life.

3

(1) Challenge Maintenance

Definition of Maintenance

P: Polish pad inside left almost must be renewed. S: Polish pad inside left nearly renewed.

Page 4: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17 4

(1) Challenge Maintenance

Maintenance Strategies Classification

Maintenance

Non Preventive Maintenance

Preventive Maintenance

Bre

akd

ow

n

Mai

nte

nan

ce

Co

rre

ctiv

e M

ain

ten

ance

Pe

rio

dic

M

ain

ten

ance

Co

nd

itio

n B

ased

M

ain

ten

ance

Pre

dic

tive

M

ain

ten

ance

Prevention Maintenance

No

M

ain

ten

ance

P: Test run OK, stopping with auto-mode very hard. S: Stopping with auto-mode not available.

Page 5: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17

• Breakdown Maintenance (BM) – Repairs or replacements performed after an equipment has failed to return to its functional state following a

malfunction or crash.

• Corrective Maintenance (CrM) – Repairs or replacements of equipment / its components with design weakness to improve reliability or

maintainability

• Operational Maintenance (OpM) – Care and minor maintenance of equipment using procedures that do not require detailed technical

knowledge of the equipment’s or system’s function and design.

• Preventive Maintenance (PM) – Maintenance strategy to retain the healthy condition of equipment and prevent failure through the

prevention of deterioration, periodic inspection, equipment condition diagnosis, or prediction algorithms. PM activities are performed before equipment fails. PM is usually performed during idle periods.

• Periodic maintenance (PeM) – Regularly scheduled (timer, counter, idle time, …) inspections are performed.

• Condition based Maintenance (CbM) – Strategy that monitors the actual condition of the equipment / its components and signals to perform

maintenance only when certain indicators show signs of decreasing performance or upcoming failures.

• Predictive Maintenance (PdM) – Prediction algorithms and methods when maintenance of in-service equipment should be performed.

Employs a surveillance system and uses measuring and analyzing data about deterioration.

• Maintenance Prevention – Strategy to operate an equipment with avoidance of any maintenance, e.g. only usage until the first

maintenance ( refrigerator, laundry machine)

5

(1) Challenge Maintenance

Maintenance Strategies Explanation

P: Machine not communicating. S: Psychiatrist hour with machine performed.

Page 6: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17

• Problem – It is not known exactly which equipment needs

when and which maintenance • “Unplanned stops” interrupting plant availability • Hard to plan (spare parts, people, production, etc.) • Equipment lifetime, plant safety, accidents with

negative impact on environment • Finally a cost problem

– The questions is "the right information in the right time“ ?

• Objectives – Minimize frequency of interruption in usage – Maximize availability and reliability – Maximize lifetime of operation / usage – Maximize production capacity and operation – Minimize total production costs – Maximize safety of workers

6

(1) Challenge Maintenance

Problem and Objectives of Maintenance

P: There's something loose in the equipment. S: We attached something in the equipment.

Preventive Maintenance

Cost

Page 7: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17

• Equipment Interfaces – Interfaces (OPC, SECS, PLC, Data Logger, DB‘s, File‘s, …) existence,

functionality, robustness, platforms, …, available – Implementation of external sensors already done in the past

• Machine Data Collection and Equipment Book – Counter-, hours of operation-, consumption-oriented – Includes disturbance and setup parameters – More difficult: setup, cycle, repair, maintenance times & statistics – Supports collection of any kind of process and product data

• Aspects and Problems – Integration standard and efficiency ( connectivity) – Configurability and flexibility of interfaces / data collection – Availability of required data, their interpretation and quality – “Fear of” observation and supervision

7

(2) Current Situation Prerequisites and Aspects

P: Note a leak on the right side. S: Note removed.

Page 8: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17 8

(2) Current Situation

Classic Manufacturing IT Systems

Within Fabs & Global …

• System Setup & Master Data management

• Extensive Data Collection Mechanisms and Data Collection Plans

• Often integrated in MES Layer

• Aggregation and Transport into global DWH of enterprise

• Mostly Analysis & Reporting at Sites

• OEE, ARAM, … KPI(s) Statistics with DWH technology

• First Sensor Integration

• First real time KPI calculation

• …

RF

ID

P: IFF does not work. S: IFF never works when it is off.

Page 9: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17

• Corrective Maintenance (CrM) is nowadays probably the most frequently used approach – but it is easy to see its limitations: – Unpredicted – Unscheduled – Low Availability – High Costs – Safety Issues

• Periodic maintenance (PeM) is probably the second often commonly used approach – Prerequisite: ODA / MDA (appr. 250 solutions only in Germany) – Often as an additional supplementing component of MES – Results inevitably in OEE, TPM, … reports and statistics – In many cases often the first step into digitalization / MES introduction – PeM decreases downtime only !

9

(2) Current Situation

Corrective and Periodic Maintenance

P: Suspect a crack in the chamber backside wall. S: Guess you're right.

Page 10: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17

• Business Models – Customer owns equipment and performs maintenance – Customer owns equipment but supplier does maintenance – Supplier owns equipment and does maintenance – And all combinations … – With or without “platform strategy” – System Integrator supports customer (more often) and / or supplier (less often)

• Aspects – Mostly in the premises of customer plant / site – Case of installation of external sensors uncertainty regarding warranty, liability – Remote access ? – Discussion about

• who has the knowledge • who owns the data • who is accountable and responsible

– No globally successful solutions in the past for “global / outsourced” maintenance platforms and virtual markets

10

(2) Current Situation

Typical Business Models

P: Machine buzzing. S: Machine reprogrammed to mumble now.

Page 11: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17 11

(3) Impact of IoT, Cloud & Big Data

Outline of IoT Boxes, Cloud Computing, Big Data

• IoT Box: IoT device and software “out of the box / in a box” with the ability to connect and integrate sensors, equipments, objects, etc. with the internet (cloud) for data exchange

• Cloud Computing: is a type of Internet-based computing that provides shared computer networks, servers, storage, applications and services to humans, systems and other devices on demand.

• Big Data: is a term for data sets that are so large or complex that traditional data processing applications are inadequate to deal with them. It include analysis, capture, data curation, search, sharing, storage, transfer, visualization, querying, updating and information privacy.

P: Mouse in the equipment. S: Cat installed.

Page 12: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17

• Integration of Things, Equipment & Systems – Isolated networks and IoT Integration everything connected, Internet technologies and tools (WebSrv) – MES, MCS, SPC, … and IoT integration of both worlds

• Data Collection, Analytics and Intelligence – Less human interaction, less effort, any and all time, small devices, invisible more seamless – Much more data distributed or centralized storage and/or processing, data stream processing & analysis

• Higher Level of Automation (automated Maintenance) – Routine scenarios and processes higher automated – Non routine work and exceptions human intelligence

• Resource Tracking, Management and Utilization – Inventory, Usage, Location, … more accurate – Additional Services and Analysis more information – For things and their behavior improved models and understanding

• Factory / Equipment Data and Status Monitoring – Much more granularity & accuracy higher precision of virtual representation of physical world – Geographical assignment improved visibility – Continuously updated less latency – In summary smaller gap between reality and virtual representation

12

(3) Impact of IoT, Cloud & Big Data

Potential contributions and advantages

P: Drive 3 is missing. S: Drive 3 found after a short search on the right tool site.

Page 13: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17 13

(3) Impact of IoT, Cloud & Big Data

Trendy Manufacturing IT Systems

• SAP Data Analytics

• IBM Watson

• TIBCO Spotfire

• Oracle Pred. Analysis

• Matlab Pred. Analysis

• SAS Pred. Analysis

• Revolution Analytics

• RapidMiner

• Azure Mach.Learning

• …

• Storage Big Data, Smart Data, Aggregated Data, … Cloud, Centralized, Distributed, …

• Analytics / Machine Learning inspect, clean, transform, model data discover, suggest, and support decision-making

• Event Stream Processor Transformation and Consolidation of data Event/data stream processing engine Real-time and complex event processing

• Event / Message Broker Event & Message stream handling capability Data forwarding with high throughput Queuing, Guarant. Delivery, Sequence, Once

• Event / Message Publisher Entity / Thing that sends events and data Many Sources: Eqp, SCADA, IoT, ERP, MES, …

• Agent Gather, transform & transport data via internet Identify actions

• Field & Sensor Gateways SW enabling the communication of devices Device data processing hub Data format: JSON (EqpID, MatID, TS, Value, UoM)

FIA (state of art)

• SAP ESP

• Oracle Stream Explorer

• TIBCO Streambase

• SAS ESP

• IBM InfoSphere

• Apache Storm

• WSO2 Siddhi

• Drools Fusion

• Azure Stream Analytics

• …

• Apache Hadoop, HBase,

Hive,

• Storm

• Disco

• Filemap

• Sphere

• MongoDB

• Cloud MapReduce

• Distributed RDB’s

• …

• [1] Data -- > Processing ?

• [2] Processing Data ?

• [3] Combination of Both ?

IoT (trend)

•Publish -

Subscribe

•Data

Streams

Event

Broker

•Complex

Event

Processor

•Stream

Processor

Event

Stream

Processor

•Aggregate

•Model

•Discover

•Visualize

•Predict

Data

Analytics

•Big Data

•Distributed

•Unstructured

•…

Data

Storage

P: Machine got very hot. S: Filled in ice cold oil.

Page 14: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17

• Today, Condition based Maintenance (CbM) is becoming increasingly popular. But due to its costs and despite of obvious advantages … – it is not yet used for less important parts of machinery. – it can be found where increased reliability, safety and efficiency of more important

resp. expensive equipment or consequences of its failure is required.

• CbM is based on using real-time data, calculating one or more indicators and, together with a set of rules, diagnose the need of maintenance. – Observing the state of the system is known as condition monitoring. Such a system

will determine the equipment's “health”, and act only when maintenance becomes necessary - according the monitored conditions.

– Development in recent years has allowed extensive instrumentation of equipment, and together with better tools for analyzing condition data, the maintenance personnel of today is - more than ever - able to decide a good time to perform maintenance on a piece of equipment.

– Ideally condition-based maintenance will support the maintenance personnel to do only the right things, minimizing spare parts and time spent on maintenance as well as increases equipment uptime !

14

(3) Impact of IoT, Cloud & Big Data

Condition Based Maintenance

P: Machine works strange. S: Machine instructed to be serious and work decently.

Page 15: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17

• Business Models – Customer owns IoT dev, Cloud, Big Data & indicates & does maintenance – Equipment Supplier owns IoT dev, Cloud, Big Data & …

• Installs on old equipment & indicates & does maintenance • sells new equipment including it & indicates & does maintenance

– Third party Supplier (e.g. SAP, AMAT, SYSTEMA) owns IoT dev, Cloud, Big Data & indicates & does maintenance

– And all combinations … – System Integrator supports customer, supplier, third party, provides services for maintenance

indications

• Aspects – The initial cost of CbM, esp. sufficient instrumentation of (already installed) equipment is high.

IoT & wireless can reduce costs. For minor equipment often higher than the equipment value. – Introducing CbM invokes a major change in how maintenance is performed, and to the whole

maintenance organization in a company. Organizational changes are in general difficult. – The technical side is not as simple: Even if measuring values as current, vibration, pressure is

easy, it is not trivial to transform them into actionable knowledge about health of equipment. – An increased number of IoT components (the CbM system itself) needs internal checking and

maintenance. – Sending all data into the cloud and storing them versus restricted capacity of the internet.

15

Upcoming Business Models

(3) Impact of IoT, Cloud & Big Data

P: Dead beetles in belt drive. S: Alive beetles in delayed delivery.

Page 16: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17 16

(4) Potential Future Solutions ?

On the Edge Computing and Machine Learning

• Edge Computing & Analytics: the ability to have intelligent embedded processing capability located as close to the sensor or device as possible and thus reduces the amount of data transmitted by isolating and sending only the pertinent data needing to be analyzed into the cloud (by several orders of magnitude). It provides IoT device access and control, localized intelligence, autonomous onboard security and privacy mechanisms.

• Machine Learning: the subject of informatics that deals with the ability of computers to learn without being

explicitly programmed. Derived from the field of pattern recognition and computational learning theory in artificial intelligence, it explores the science of algorithms that can learn from and make predictions on data.

P: Temperature sensor delivers bad data. S: Temperature sensor inverted to deliver good data.

Page 17: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17

• Evolution of Servers – Numerous commercial “IoT edge server” boxes act as localized intelligent controllers and processors.

• Essentially modified PC servers may become “middleware solutions” to deploy analytics at the edge • Who : Intel, Dell, HPE, Cisco, McAfee, SAP, … • What: lightweight (database platform , Parstream, Watson Capabilities, Connected Analytics, mashery (APIs), IoT Developer

Kit, IoT Gateway, Access NetFront Browser, JavaScript engines based on node.js or PhantomJS, Hana (synch of data between enterprise and remote locations), …

– E.g. Camera: device’s role changes from video streaming to intrusion detection - much higher value capability (event generator) !

• Evolution of Analytics and Machine Learning

– Creating an analytics model in one location • Creating the analytics model involves: collecting data, storing data, preparing the data for analytics (some ETL functions),

choosing the analytics algorithm, training the algorithms, validating the analytic goodness of fit etc. The output of this trained model will be rules, recommendations, scores, etc .

• Only then, can we deploy this model.

– Deploying and executing the analytics model at multiple points • Predictive Model Markup Language (PMML) is an XML-based predictive model interchange format and becomes important

for the ability to deploy models in multiple locations. • It provides a way for analytic applications to describe and exchange predictive models produced by data mining and machine

learning algorithms. • It supports common models such as logistic regression and feed forward neural networks.

– Decentralized processing with inherent complexity • Known: management and replication of master data, security, storage, program code, etc. • New aspect: the process of creating the analytics model on one machine and deploying it on another machine

17

(4) Potential Future Solutions ?

Trends: Distributed Processing and Analytics / Learning

P: Machine needs maintenance. S: Machine needs operator.

Page 18: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17 18

(4) Potential Future Solutions ?

Future Manufacturing IT Systems ?

1

Fab M

essage Bu

s (Mo

M)

EDS 1

Junior 1.1

Junior 1.16

Junior n.1

Junior n.16

Electrical

Cabinet

Electricity Sen

sors

Electrical C

on

nectio

n

EDS 2

Junior 2.x

EDS n

TalkingEn

ergy Interface

Oth

er Sen

sors

BI / RI Suite

MES OEE/RTC

MES EDC

Auto-mation

OPC UA ?

OPC UA ?

OPC UA ?

ESP ? Pattern Recog. Event Gener.

SYS-TE App @Server - Industrial App (OEE Mon, Pred.Maint, Exceptions, …)

Cloud

Server

2

3

4

Local MES

Site / Cust. 1

Local MES

Site / Cust. 2 HADOOP ?

Info & Feed

Back to Cust.

MES in the Cloud ?

MES Core

one

example of

many

Lean Automation Box

P: Auto-Delivery failed. S: Manual Delivery worked.

Page 19: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17

• Predictive Maintenance (PrM) is a change in paradigm and will roll over the maintenance strategies – PrM is a core use case of industry 4.0; supported by many R&D tenders and

programs – PrM promises big cost reductions and efficiency wins; therefore many customers,

vendors, suppliers, etc. are working on it

• PrM is an extended and distributed CbM – The main promise of predictive maintenance is to allow convenient scheduling of

corrective / preventive maintenance, and to prevent unexpected equipment failures.

– The key is "the right information in the right time". – Main objective:

• Maximizes uptime !! • with minimal disruption of normal system operations • while allowing for budgeted and scheduled repairs

– Prediction is a tough business: Brexit, Trump election, Galaxy Note 7 kills iPhone

19

(4) Potential Future Solutions ?

Predictive Maintenance Approaches

P: ?

S: !

Page 20: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17

• New Service Models – Software on Demand (Application Service Provider) – Infrastructure as a Service (IaaS), Platform as a Service (PaaS), Software as a Service (SaaS), .... – IoT Edge Devices and Boxes with pre-installed and pre-configured middleware, interfaces, software packages – Services for Data Analytics and Machine Learning

• Business Models – Customer owns IoT dev, Cloud, Big Data & Edge Solutions & indicates & does maintenance – Equipment Supplier owns IoT dev, Cloud, Big Data & Edge Solutions & …

• Installs on old equipment & indicates & does maintenance • sells new equipment including it & indicates & does maintenance

– Third party Supplier (e.g. SAP, AMAT, SYSTEMA) owns IoT dev, Cloud, Big Data & Edge Solutions & indicates & does maintenance

– And all combinations … – System Integrator supports customer, supplier, third party, edge solution, provides services for maintenance

indications

• Aspects – PrM even expands the complexity and aspects of CBM – Many competing platforms and techniques, no standard yet – if at all – Who (Customer, OEM, Supplier, Third Party, …) can contribute what ? – Security, Privacy, Data Ownership, Legal Aspects, Autonomous, …

20

(4) Potential Future Solutions ? Future Business Models and System Architectures

Page 21: Predictive Maintenance from System Integration Perspective · Manfred Austen CEO Predictive Maintenance from System Integration Perspective My Survey on Current Situation to Potential

© 2017 SYSTEMA GmbH / 19.01.17

• A company should establish a requirement list for PrM including questions such as: – In which environment (systems, equipment, machines, tools, …) must PrM work ? – Does the environment (systems, equipment, machines, tools, …) meet the pre-requisites? – To what extent can an initial test system be extended to other devices and machines? – Is the solution scalable? – For which systems, devices, machines, tools, is a PrM solution useful? – Where are risks for downtime and unforeseen failures? – Where and how can systems, devices, machines, tools, be accessed and what data do we have

to collect or generate? – How is data analysis done? – Do we want to use PrM as a cloud service and where is the cloud hosted? – Which backend system is the right one for us? – Which failures can also be rectified remotely? When and under which conditions must a

technician be onsite? – Which provider can support us with its industry expertise? – Does the case (ROI) pay off?

• Recommendation: Start to build and follow your own strategy !

21

(5) Conclusion

PrM questionnaire and recommendation